PROJECT SUMMARY/ABSTRACT Medication nonadherence is common and leads to adverse outcomes and increased healthcare costs. Many studies have evaluated patient or medication factors that predict poor adherence. However, despite the crucial role of the healthcare provider in prescribing and monitoring therapy, few studies have focused on continuity of care or the provider's characteristics as possible determinants of adherence. The goal of this work is to understand medication adherence in Medicare Part D beneficiaries and to evaluate the role of the provider in adherence. An equally important goal of this career development award is to allow Dr. Holly Holmes to gain expertise in geriatric pharmacoepidemiology in order to become an independent investigator in medication- related problems in older cancer patients. Dr. Holmes is a geriatrician and prior pharmacist, and will achieve her goals by completing an MS degree, specialized coursework, database skills training, and regular seminars, meetings, and workshops. She will accomplish her research and training plan with the guidance of a highly qualified team of mentors. Nonadherence will be examined in two clinical situations: hypertension and breast cancer. First, hypertension will be evaluated to be able to understand adherence in a well-studied disease state and incorporate the innovative approach of evaluating the role of the provider and continuity of care. An analogous approach will then be applied to nonadherence to oral hormonal anticancer therapies. Nonadherence will be studied in Part D beneficiaries, which is a new area, with increasing amounts of high quality pharmacy data. The specific aims are: 1) to determine whether patient and medication factors that predict nonadherence in other populations predict nonadherence among persons enrolled in Medicare Part D, 2) to determine whether lower continuity of care predicts medication nonadherence, and 3) to quantify the extent to which nonadherence is predicted by provider characteristics. These aims will be accomplished in a multilevel model using patient, medication, and provider variables from Medicare claims and Part D claims files. By including continuity of care and the provider in the analysis, the proposed research will provide a more comprehensive understanding of factors that impact medication adherence. In addition, this research will provide new and useful information about current adherence patterns to oral hormonal anticancer therapies that could help model treatment effects and design intervention trials. By achieving our aims, we may identify new ways to approach future studies of medication adherence, including studies of the effects of nonadherence on cancer-specific outcomes. This study will provide the candidate with valuable expertise in the use of claims data, Part D data, and cancer registry data. The research and career development plan will help the candidate develop into an independent investigator.